Marker Substrates. CLint values were obtained for each of the prototypic substrates, tolbutamide, diazepam, metoprolol, ibuprofen, propranolol, dextromethorphan, omeprazole, diltiazem, testosterone, and verapamil in three individual preparations of pooled HLM. Table 3 compares the CLint for the three batches of HLM together with the mean and values obtained from the literature. For substrates with significant CYP3A4 metabolism, diltiazem, testosterone, and verapamil, CLint was significantly higher in batch 1 compared with batches 2 and 3. The coefficient of variation was generally int determined from one pool of HLM.
Substrates were incubated with recombinant CYP1A2, -2C9, -2C19, -2D6, and -3A4, respectively, using the RSP as described under Materials and Methods
. Figure 1
displays the loss of propranolol against time by the five different CYP isoforms used in the automated assay and shows significant metabolism by CYP1A2, -2C19, and -2D6. Table 4
shows the CYP CLint
of individual CYP isoforms to oxidative metabolism for each marker substrate. The range of CYP CLint
determined was 0.03 to 7 µl · min1
pmol of P4501
. The percentage contributions of individual CYPs toward oxidative metabolism of a compound in HLM were estimated, and Table 5
compares our values with common literature assignments. Tolbutamide (CLint
determined by Vmax
) was metabolized by both CYP2C9 (70%) and CYP2C19 (30%), diazepam (10 µM) by CYP2C19 (100%), ibuprofen (10
µM) by CYP2C9 (90%) and CYP2C19 (10%), and omeprazole (3 µM) by CYP2C19 (68%) and CYP3A4 (32%), respectively. Metoprolol (3 µM) and dextromethorphan (3 µM) are primarily CYP2D6 substrates and propranolol (3 µM) was metabolized by CYP2D6 (59%), CYP1A2 (26%), and CYP2C19 (15%). Diltiazem (3 µM), testosterone (10 µM), and verapamil (3 µM) were predominantly metabolized by CYP3A4.
For each compound, the sum of the CLint
(µl · min1
) from the five individual isoforms was compared with the respective CLint
derived from the mean of three separate HLM pools (Table 6
). Figure 2
shows the correlation (r2
= 0.8, P
int and HLM CLint
. Compounds with a HLM CLint
may be described as low clearance, 8 to 65 µl · min1
as intermediate, and >65 µl · min1
as high clearance.
Predictions of HLM CLint
from the sum of individual CYP CLint
were excellent for tolbutamide (CYP CLint
= 1.3 µl · min1
versus HLM CLint
= 0.7 µl · min1
), diazepam (2 ± 1 versus 3 ± 1), and metoprolol (7 ± 0 versus 6 ± 1). All predictions of HLM CLint
from the individually summed CYP CLint
except for those derived for propranolol (CYP CLint
= 55 ± 15 µl · min1
versus HLM CLint
= 15 ± 0 µl · min1
) and omeprazole (131 ± 25 versus 34 ± 14) were within 3-fold. The summed CYP CLint
of these two compounds significantly overestimated their CLint
determined in HLM.
To investigate the relationship between the CLint of propranolol with increasing HLM protein concentration, CLint was determined at 0.4, 1, and 2 mg · ml1 of HLM. Figure 3 shows the relationship between increasing microsomal protein and decreasing CLint of propranolol.
The metabolite profile for the CYP-dependent clearance of dextromethorphan observed by HPLC-mass spectrometry was as predicted from Von Moltke et al. (1998), where microsomes containing individual CYPs expressed by a human lymphoblastoid expression system were used (Fig. 4). Dextromethorphan was chosen for this detailed analysis, because four of the five major human CYPs (CYP2C9, -2C19, -2D6, and -3A4) have been implicated in its metabolism. Dextromethorphan was incubated at 30 µM with all five isoforms, and as expected, two metabolites, dextrorphan and 3-methoxymorphinan, were observed, based on their m/z values and distinguished using standards. Based on UV response, 88% of the metabolites formed from dextromethorphan (m/z = 272) were dextrorphan (m/z = 258) and 12% were 3-methoxymorphinan (m/z = 258), which compares well with Von Moltke et al. (1998) (98 and 2%, respectively). The isoform responsible for dextrorphan formation was primarily CYP2D6 (92% versus 97%; as determined from Von Moltke et al., 1998) with minor contributions from CYP2C9, -2C19, and -3A4 ( for 3-methoxymorphinan formation were CYP2C9 (43% versus 55%), CYP3A4 (42% versus 20%), CYP2C19 (8% versus 16%), and CYP2D6 (7% versus 9%). In addition, it was also determined that CYP1A2 metabolized propranolol (m/z = 260) to the expected N-deisopropylation product (m/z = 218) (Yoshimoto et al., 1995) and CYP2D6 metabolized propranolol to the expected hydroxylated product (m/z = 276), although the regiochemistry of hydroxylation was not investigated. Several other markers also generated the product profile as expected from the literature (data not shown).